Q. My 24-year-old daughter was diagnosed with bipolar disorder in college. She does not take any medications and so far has done well. Recently, she developed Bells Palsy and was prescribed a course of Prednisone treatment by a neurologist. She listed the possible bipolar diagnosis on the doctor's information sheet. Within a week, she ended up in the emergency room and was diagnosed with steroid-induced psychosis.

Two weeks after stopping the Prednisone, she was still hypomanic. She started to feel better two days ago. Is there a greater risk of a negative reaction to steroid treatment in a patient with a preexisting mood disorder?

A. Good question--the basic answer is, "It's not clear." Let's back up for a minute, though. Your daughter apparently experienced a reaction that is seen in roughly 6% of patients treated with steroids. With respect to Prednisone, there is clearly a dose-relationship: fewer than 2% of patients who get less than 40 mg/day develop psychiatric symptoms, versus 5% with doses of 40-80 mg/day and over 18% with doses >80 mg/day.

Reactions usually begin within first 1-3 weeks of treatment, but may require 4-6 weeks to resolve after cessation of steroids (Hall et al, 1979). Reus & Wolkowitz (In Psychoneuroendocrinology, edited by Wolkowitz and Rothschild, APA, 2003) observe that "Previous history of psychiatric disorder has?not been shown to reliably relate to corticosteroid-induced mental change, although it is quite possible that it contributes to shaping the symptom picture once a drug-induced effect occurs?"

Symptoms usually remit promptly with discontinuation of the drug, but delayed onset and more prolonged courses have been reported [Pies, 1981]. Female gender and the presence of autoimmune disease are associated with increased risk of steroid-related psychiatric problems, but the gender finding may be related to the higher incidence of autoimmune disease (such as lupus) in females. Certainly, caution is warranted in starting bipolar individuals on steroids, since higher doses pose some risk of destabilizing their illness; however, the data do not show conclusively that bipolar disorder is reliably associated with a high risk of steroid-induced psychosis.

January 2004

Disclaimer Back to Ask the Expert